مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 Min (Max)-CS Modules I. M. A. Hadi, R . N. Majeed Department of Mathematics, College of Ibn-Al-Haitham , University of Baghdad Received in:25 August 2011, Accepted in:20 September 2011 Abstract. In this paper, we give a comprehensive study of min (max)-CS modules such as a closed submodule of min-CS module is min-CS. Amongst other results we show that a direct summand of min (max)-CS module is min (max)-CS module. One of interested theorems in this paper is, if R is a nonsingular ring then R is a max-CS ring if and only if R is a min-CS ring. Key words: CS-module, min-CS module, max-CS module, uniform-CS module. 1- Introduction Throughout the paper all rings R are commutative with identity and all R-modules are unitary. We write A  M and A  e M to indicate that A is a submodule of M and A is an essential submodule of M, respectively. Recall that anR-module M is called an extending module (or, CS-module) if every submodule is essential in a direct summand of M or M is extending if and only if every closed submodule is a direct summand, [1, p.55]. In this paper definitions, notations, examples and fundamental results of min (max)-CS modules are introduced. 1.1 Definition: [2] An R-module M is called min-CS module if every minimal closed submodule of M is a direct summand of M. A ring R is called min-CS if it is min-CS R-module. 1.2 Definition: [2] An R-module M is called max-CS module if every maximal closed submodule of M with nonzero annihilator is a direct summand of M. A ring R is max-CS if it is max-CS R-module. Recall that an R-module M is -injective (quasi-continuous) if and only if M satisfies C1 (M is extending) and C3, where M is said to satisfy the C3 if the sum of any two direct summands of M with zero intersection is a direct summand of M. [3, p.18] 1.3 Remarks and Examples 1. Every CS-module is min-CS and max-CS. Proof: It follows directly by [1, p.55]. 2. Every semisimple module is max-CS and min-CS. In particular ℤ2, ℤ3, ℤ6, ℤ10,…, ℤ30 as a ℤ-module is max-CS and min-CS. Proof: By [1, p.55], every semisimple module is CS. Hence the result follows by remark 1. مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 3. Every uniform R-module M is min-CS and max-CS. In particular each of ℤ-module ℤ, ℤ4, ℤ8, ℤ9, ℤ16 is min-CS and max-CS. 4. If R is a semisimple ring, then every R-module M is injective, by [4, theorem 1.18, p.29]. Hence M is max (min)-CS module since every injective module is CS. By [1, p.16]. 5. The ℤ-module ℤ12 is max-CS and min-CS. The submodules of ℤ12 are 2 , 3 , 4 , 6 , 0          and ℤ12. Since 123 4     ¢ . So each of 3  and 4  are direct summands. Hence they are closed submodules. 2   e ℤ12 and e6 3     imply hat 2  and 6  are not closed. Thus M is CS and so max-CS and min-CS. 6. It is easy to check that each of the ℤ-modules ℤ18 and ℤ24 are min-CS and max-CS. 7. Let M be a module whose lattice of submodules is the following: 2 1 2 M N N N 0     / \ \ / It is clear that N1 is closed in M, but it is not a direct summand of M. So M is not CS. Also N1 is a minimal closed submodule of M. Hence M is not a min-CS module. Notice that N1  N2  e M, so it is not closed submodule of M. It follows that N1 is a max-closed submodule of M. Hence M is not a max-CS module. 8. Every -injective is min-CS and max-CS. Proof: It follows by the definition of -injective module and remark 1.3 (1). 9. Let M be the ℤ-module ℤ8  ℤ2. M is not CS-module, since there exists a submodule N = { (2 ,1 ), (4, 0), (6, 1 ), (0,0) } which is closed but not a direct summand. Moreover N is minimal closed, so M is not min-CS module. On the other hand M is a max-CS module, since the only maximal closed submodules of M are ℤ8  ( 0 ) and < (3,1) >, and (ℤ8  ( 0 ))  (( 0 ) ℤ2) = M and < (3,1) >  (( 0 )  ℤ2) = M . Hence we deduce that M is max-CS. 10. Let M 1 and M 2 be two R-modules such that M 1 is isomorphic to M 2 (M 1≅M 2), then M1 min (max)-CS if and only if M2 is min (max)-CS. Mahmoud A.Kamal and Amany M.Menshawy in [5] gave the following: An R-module M is called min-CS if every simple submodule of M is essential in a direct summand. However this definition of min-CS is different from definition 1.1, since the ℤ-module M = ℤ8  ℤ2 is not min-CS in our sense. However it is min-CS (in sense of Kamal and Menshawy) since ℤ8 is CS, so min-CS (in sense of Kamal and Menshawy) and ℤ2 is simple, so semisimple. مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 Hence by applying [5, lemma 3, p.166], M is min-CS (in sense of Kamal and Menshawy). 1.4 Proposition: Let M be an R-module, and let I be an ideal of R such that I ⊆ annM. M is max- CS R-module then M is max-CS (R/I)-module and the converse is true if annM ≠ annN, for all maximal closed submodule N ≨ M. Proof: Let N be a maximal-closed (R/I)-submodule of M and annR/IN ≠ 0R/I = I. It is easy to see that N is a maximal closed submodule of M. Since annR/IN ≠ I = 0R/I, so there exists r + I  R/I with rI such that r + I  annN, hence r ≠ 0 and rN = 0. Thus annRN ≠ 0 and so that N is a direct summand of M. Conversely, let N be a maximal closed R-submodule of M with annN ≠ 0. Hence N is a maximal closed (R/I)-submodule. Now, since ann M  ann N, there exists r  ann N and r  ann M. Thus r  I; that is 0R/I = I ≠ r + I and (r + I)N = rN = 0. Hence ann R/IN ≠ 0R/I. But M is a max-CS (R/ I)-module, so N is a direct summand. 1.5 Proposition: Let M be an R-module, let I be an ideal of R such that I  ann M. Then M is min-CS R- module if and only if M is min-CS (R/ I)-module. Proof: It is straight forward, so it is omitted. Recall that an R-module M is called a uniform extending (or uniform-CS) if every uniform submodule is essential in a direct summand. [1, p.55]. Al-Hazmi in [2,p.24], mentioned that min-CS and uniform-CS are equivalent concepts without proof. We shall prove this equivalence, but first we need the following lemmas. 1.6 Lemma: Let N be a submodule of an R-module M. N is minimal closed if and only if N is uniform-closed (that is every closed submodule of N is essential in N). Proof: () It is enough to prove that N is uniform Let V, W be two nonzero submodules of N. Suppose V  W = (0). Hence, there exists V'  N such that V' is a relative complement of V and hence V' is closed in N. Since N is minimal closed of M thus N is closed in M, so by [4, proposition 1.5, p.18], V' is closed in M and 0 ≠V'  N. Thus N is not minimal closed submodule of M, which is a contradiction. Therefore, N is uniform-closed submodule. () Suppose that there exists a closed submodule V of M such that V  N. But N is uniform. So V  e N. Hence V = N, since V is closed. Thus N is minimal closed. 1.7 Lemma: If U is a uniform submodule of M such that U e K  M. Then K is uniform. Proof: مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 Let V and W be two nonzero submodules of K. Since U  e K, U  W ≠ (0) and U  V ≠ (0). But U is uniform submodule of M. So (U  V)  (U  W) ≠ (0). Hence U  (V  W) ≠ (0). Thus V  W ≠ (0); that is K is uniform. 1.8 Lemma: Let U be a submodule of an R-module M. Then U is uniform closed if and only if U is maximal uniform; that is U is maximal in the collection of uniform submodules of M . Proof: () Suppose there exists a uniform submodule V of M such that U  V. Since V is uniform so U  e V. But U is closed so U = V. Thus U is a maximal-uniform submodule. () It is enough to show that U is closed. Suppose there exists V  M such that U  e V. Hence by lemma1.7, V is uniform. Thus U = V, since U is maximal uniform, so that U is closed. By combining lemma 1.6 and lemma 1.8, we have the following: 1.9 Corollary: Let U be a submodule of an R-module. Then the following are equivalent: (1) U is minimal-closed. (2) U is uniform-closed. (3) U is maximal-uniform. Now, we can give the proof of the following result. 1.10 Proposition: [2, p.24] Let M be an R-module. M is uniform-CS if and only if M is min-CS. Proof: () Let U be a maximal uniform submodule of M. Since M is uniform-CS, so U is essential in a direct summand V. Then by lemma 1.7, V is uniform. Hence U = V since U is a maximal uniform submodule. Thus U is a direct summand of M, it follows that every minimal-closed is a direct summand by lemma 1.6. So that M is a min-CS module. () Let U be a uniform submodule of M. By [4, Exc.13, p.20], there exists a closed submodule V of M such that U  e V. Hence by lemma 1.7, V is uniform. Thus V is a closed-uniform submodule of M, and hence by lemma 1.6, V is minimal-closed. So that V is a direct summand of M, since M is min-CS module. It follows that U is essential in a direct summand. Thus M is uniform-CS. The following result is given in [2, lemma 3.1.1, p.45], we give the details of proof. 1.11 Proposition: Let M be an indecomposable R-module with a uniform submodule. If M is a min-CS module, then M is uniform. Proof: مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 By hypothesis M has a uniform submodule, say U. By [4, Exc.13, p.20], there exists a closed submodule K of M such that U  e K. Hence by lemma 1.7, K is a uniform submodule of M. Let C = {K: K is a uniform submodule of M and U  K}. Hence C ≠  , and so that by Zorn's lemma C has a maximal element say W. It is clear that W is a maximal-uniform. So it is a minimal-closed submodule of M. Thus W is a direct summand of M, since M is a min-CS module. Then W  V = M for some submodule V  M. But M is indecomposable, hence V = (0). Thus W = M ; that is M is uniform. 1.12 Corollary: Let M be an indecomposable R-module with a uniform submodule. Then M is min-CS module if and only if M is a uniform module. 1.13 Proposition: Let M be an R-module. The following statements are equivalent for a module M:- (1) M is a min-CS module. (2) For every minimal-closed submodule A of M, there is a decomposition M = M 1  M 2 such that A is a submodule of M1 and M2 is a complement of A in M. Proof: (1)  (2) Let A be a minimal-closed submodule in M. Therefore A is a direct summand of M since M is a min-CS module. That is M = A  M' for some submodule M' of M. It is clear that A is a submodule of A, and it is easy to check that M' is a complement of A. (2)  (1) To prove M is a min-CS module. Let A be a minimal-closed submodule of M. Therefore, there is a decomposition M = M 1  M 2, where M1 and M2 are two submodules of M and A is a submodule of M1 and M2 is a complement of A in M. Since M2 is a complement of A in M, then A  M 2  e M. But A is a closed submodule in M and A is a submodule in M 1, therefore A is closed in M1. Therefore, A  M 2 is closed in M1  M 2 = M by [4, Exc.15, p.20]. Thus A  M 2 = M . So that A is a direct summand in M. Hence M is a min-CS module. 1.14 Proposition: Let M be a min-CS R-module. If N is a closed submodule, then N is a min-CS module. Proof: Let U be a minimal closed submodule of N. Since N is a closed submodule in M, then by [4, proposition 1.5, p.18], U is closed in M. We claim that U is a minimal closed submodule of M. To prove our assertion: Suppose there exists a closed submodule V of M such that V  U. But V  U  N, implies V is a closed submodule in N by [4, p.18]. But U is a minimal closed submodule of N so U = V. Thus U is a minimal closed submodule of M, and hence U is a direct summand of M, since M is a min-CS module. Hence M = U  W for some W  M. It follows that N = (U  W)  N and by modular law, N = U  (W  N); that is U is a direct summand of N and so N is a min-CS module. مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 1.15 Remark: The converse of the previous proposition is not true in general For example: The ℤ-module ℤ8  ℤ2 is not min-CS, but N = ℤ8  (0) ≅ ℤ8 is min-CS. 1.16 Corollary: A direct summand of min-CS module is min-CS. Proof: Let M be a min-CS module and let N be a direct summand of M. Hence by [4, Exc.3, p.19], N is a closed submodule of M. Hence, N is a min-CS module by proposition 14. 1.17 Corollary: Let M be an R-module. If M  M is min-CS, then M is a min-CS module. Proof: It follows by corollary 1.16. Corollary 1.16, lead us to give the following example:- Let M = ℤ8  ℤ2  ℤ3 as a ℤ-module. Thus M is not min-CS because if it is, then N = ℤ8  ℤ2  M is min-CS by corollary 1.16, which is a contradiction. 1.18 Remark: The condition N is closed can not be dropped from proposition 1.14 as the following example shows:- Let M be the ℤ-module ℤ16  ℤ2. M is min-CS. Let N = ( 2 )  ℤ2. It is clear that N  e M, so N is not closed in M, also N is isomorphic to ℤ8  ℤ2 which is not a min-CS module. 1.19 Proposition: Let M be a finitely generated or multiplication R-module. Then if every maximal submodule is a direct summand, then M is a max-CS module. Proof: Let A be a maximal closed submodule of M. Since M is a finitely generated or multiplication module. Then by [6, theorem 2.3.11, p.28] and by [7, theorem 2.5 (1)], there exists a maximal submodule B of M such that A  B. But B is a direct summand, by hypothesis; hence B is a closed submodule by [4, Exc.3, p.19]. It follows that A = B. Thus A is a direct summand of M. Hence M is a max-CS module. The converse of the previous proposition is not true in general, as the following example shows: 1.20 Example: ℤ as a ℤ-module is a max-CS module; also ℤ is a multiplication ℤ-module. مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 But 2ℤ as a ℤ-module is a maximal submodule of ℤ and it is not a direct summand. Note that we have an analogous result to corollary 1.16, for max-CS modules, that is a direct summand of max-CS module is max-CS module, but first we prove the following: 1.21 Proposition: Let M be an R-module, and let N be a closed submodule of M. If M is a max-CS module, then (M/N) is a max-CS R-module, provided M is not faithful. Proof: Let (B/N) be a maximal closed submodule in (M/N), with annR(B/N)≠ 0. We claim that B is a maximal closed submodule in M. To prove our assertation: First, assume that B  e L  M. But N is closed in M and N  B  L, so N is closed in L. Hence by [4, proposition 1.4, p.18], N  B  e L implies B/N  e L/N  M/N. Hence (B/N) = (L/N), since (B/N) closed in (M/N). Thus B = L and B is closed in M. Now, assume there exists a closed submodule B' of M such that B  B', hence N  B'  M. Then, by [4, Exc.16, p.20], (B'/N) is closed in (M/N) and so that (B/N)  (B'/N). This implies (B/N) = (B'/N), since (B/N) is a maximal closed submodule in (M/N). Moreover annRB  annRM ≠ 0, so annB ≠ 0. Since M is max-CS then B  K = M for some K  M, and hence (B/N)((K+ N)/N) = (M/N). It follows that (M /N) is a max-CS module. 1.22 Corollary: A direct summand of a max-CS R-module M is a max-CS module, provided M is not faithful. Proof: Since N is a direct summand of M, so M = N  W for some W  M. Hence (M/W) isomorphic to N by second isomorphism theorem. But (M/W) is a max-CS by proposition 1.21. So N is a max-CS module, by remark 1.3 (10). 1.23 Corollary: Let M be an R-module. If M  M is a max-CS module. Then M is max-CS module provided M is not faithful. Proof: It follows by corollary 1.22. Recall that a proper submodule N of an R-module M is called prime submodule if whenever r  R, x  M, r x  N implies x  N or r  [N:M]. An R-module M is called a prime module if annRM =annRN for each submodule N of M . Equivalently, M is called a prime module if (0) is a prime submodule of M, [8]. 1.24 Corollary: Let M be a not faithful prime max-CS R-module. If f : M  M' is an epimorphism, then M' is a max-CS module. Proof: M is a max-CS module. Then (M/ker f) ≅ M', by the 1 st fundamental theorem. We claim that ker f is clsed submodule in M. To show this, let ker f ≨ e L  M. Assume there exists 0 ≠ x  L such that x  ker f. مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 Then there exists 0 ≠ r  R such that 0 ≠ r x  ker f. Hence rf(x) = 0. Since M is a prime module, so either f(x) = 0 or r  annM. But f(x) ≠ 0 since x  ker f, hence r  annM and this implies r x = 0 which is a contradiction. Thus ker f = L and so that ker f is a closed submodule of M. Since annM ≠ 0 by hypothesis. Then by proposition 1.21, (M/ker f) is a max-CS module and hence by remark 1.3 (10) M' is a max-CS module. 1.25 Corollary: Let M be a max-CS not faithful R-module. If N is a prime submodule of M and [N R : M] = annRM. Then (M/N) is a max-CS R-module. Proof: Since N is a prime R-submodule of M, then (M/N) is a torsion free (R/[N:M])-module, [8, p.61-69]. Hence (M/N) is a torsion free (R/annM)-module, and so that N is closed (R/annM)-module of M by [9, remark 3.3,p.48]. It follows that N is closed R-submodule of M. Then by proposition 1.15, (M/N) is a max-CS R-module. 1.26 Corollary: If M is a max-CS not faithful R-module and N is a submodule of M, such that (M /N) is torsion free, then (M/N) is max-CS. Proof: Since (M/N) is torsion free R-module. Then N is closed submodule in M by [9, remark 3.3, p.48]. Hence the result follows by proposition 1.21. Now we have the following note for min-CS modules. 1.27 Remark: The homomorphic image of min-CS need not be a min-CS, as the following examples show. 1.28 Examples: (1) Let M be the ℤ-module ℤℤ. It is clear that M is a min-CS module. Let N = 8ℤ2ℤ, and let : M  (M/N) be the natural projection. Since (M/N) is isomorphic to ℤ8ℤ2, so (M /N) = (M) is not min-CS module. (2) Let M be the ℤ-module ℤℤ2. M is min-CS. Let N = 8ℤ( 0 ), N  e ℤ( 0 ), but (M/N) = ℤ8ℤ2 is not min-CS. (3) Let M = ℤ8ℤ be a ℤ-module, M is a CS-module. So M is a min-CS module. Let N = (0) (2) . (M/N) is isomorphic to ℤ8ℤ2 which is not min-CS module. 1.29 Theorem: Let M be an R-module. If M is a faithful, finitely generating and multiplication R-module, then M is a max-CS module if and only if R is a max-CS ring. Proof: () If M is a max-CS R-module. Let I be a maximal closed ideal of R such that ann I≠0. We claim that N = IM is a maximal closed R-submodule of M. First of all N = [IM :M]M since M is multiplication and by [10, proposition 3.31] [IM:M] is a closed ideal of R. But M is finitely generating faithful multiplication, then by [7, theorem 3.1] I = [IM:M]. مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 Thus by [10, proposition 3.31, ch.3], N = [IM:M]M is a closed submodule of M. Now, to prove that N is a maximal submodule of M. Suppose there exists a closed submodule W of M such that N  W. Hence N = IM = [N:M]M  [W:M]M = W and by [7, theorem 3.1], I = [N:M]  [W:M]. On the other hand by [10, proposition 3.31, chapter three], I = [N:M] and [W:M] are closed ideals of R. Hence I = [W:M], since I is a maximal closed ideal of R. It follows N = W and N is maximal closed submodule of M. But M is faithful multiplication, so annRN = annRI ≠ 0. Thus N is a direct summand of M. So that N  L = M for some L  M. But N = IM, L = [L:M]M, so that IM  [L:M]M = M and hence (I  [L:M])M = M . But by [7, theorem 1.6] (I  [L:M])M = IM  [L:M]M = (0), and so (I  [L:M])M = 0; that is (I  [L:M])  annM = (0). Thus I  [L:M] = (0), so that M = RM = (I  [L:M])M and hence R = I  [W:M] by [7, theorem 6.1]. () Let N be a maximal closed submodule of M with annN ≠ 0. Since N is closed, then by [10, proposition 3.1] [N:M] is closed in R. We claim that [N:M] is a maximal closed ideal in R. To show this, assume J is a closed ideal in R such that [N:M]  J, hence N = [N:M]M  JM . But JM = [JM:M]M and by [7, theorem 6.1] [JM:M] = J and by [10, proposition 3.1] JM is a closed submodule of M. It follows that N = [N:M]M = JM , since N is a maximal closed submodule of M. Then by [7, theorem 3.1], [N:M] = J; that [N:M] is a maximal closed. But annN = ann[N:M] since M is faithful multiplication. Thus [N:M] is a maximal closed ideal of R, with annR[N:M] ≠ 0. On the other hand R is a max-CS ring, so [N:M] is a direct summand of R. Thus R = [N:M]  T where T is an ideal of R, and hence M = RM = ([N:M]  T)M = [N:M]M  TM. But by [7, theorem 1.6], [N:M]M  TM = ([N:M]  T)M = 0M = 0 So M = [N:M]M  TM; that is M = N  TM. By the same argument of proof of theorem 1.29, we have the following: 1.30 Proposition: Let M be a faithful, multiplication and finitely generated R-module, then M is a min-CS ring if and only if R is min-CS ring. Next we have for nonsingular rings, the concepts min-CS ring and max-CS ring are equivalent. But first we need the following which is given in [2, lemma 2.1.1, p.31]. We give the details of proof. 1.31 Lemma: For a ring R, a complement of minimal (maximal) closed ideal of R is a maximal (minimal) closed ideal of R. Proof: () Let I be a minimal closed ideal of R. Let J be the relative complement of I. Then by [4, proposition 1.3, p.17], (I  J)  e R, and J is closed in R, by [4, proposition 1.4, p.18]. Now, we shall show that J is a maximal closed ideal in R. Assume that there exists a closed ideal J* in R, such that J  J*   R. It follows that J*  I ≠ (0) because J is the largest ideal in R such that I  J = (0). مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 On the other hand I is a minimal closed ideal in R, so it is uniform closed by lemma 1.6. Hence J* I  e I. But J  e I, hence (J* I)  J  e I  J  e R and so that (J* I)  J  e R, by [4, proposition 1.1(a), p.16]. But J* I  J* and J < J*, hence (J* I)  J  J*. Thus J*  e R, which is a contradiction. () If T is a maximal closed ideal of R and V its complement. To prove V is minimal closed in R, it is enough to show that V is uniform by lemma 1.6. Assume A  B = (0) for some two nonzero ideals A and B of V. Now, there exists a closed ideal I in R such that A  T  e I, by [4, Exc.13, p.20]. Hence T   I which is a contradiction, since I is closed in R and T is a maximal closed ideal in R. Thus V is uniform closed, that is V is minimal closed. 1.32 Note: By a similar argument of proof of lemma 1.31, we get analogous results for submodules. Recall that a ring R is semiprime if for each x  R, x 2 = 0, then x = 0, [4, p .2]. Now, we can give the following theorem: 1.33 Theorem: Let R be a nonsingular ring. Then R is max-CS if and only if R is min-CS. Proof: () If R is a max-CS ring. Let I be a minimal closed ideal in R. By [4, theorem 2.38, p.65], I = ann annI, hence annI ≠ 0, but R is nonsingular, so R is semiprime by [4, proposition 1.27(b), p.35], which implies I  annI = 0. Let J be the relative complement of I, so by lemma 1.31, J is a maximal closed ideal in R. Also by [4, theorem 2.38, p.65], J = ann annJ, so annJ ≠ 0. Therefore J is a direct summand of R since R is max-CS. It follows that J = for some idempotent e in R. On the other hand, IJ  I  J = (0), implies that J  annI. But I  annI = (0), so J = annI since J is the largest ideal in R such that I  J = (0). Thus annJ = ann annI = I and so < 1 – e > = I. It follows that I is a direct summand of R. () If R is a min-CS ring. Let I be a maximal closed ideal in R, with annI ≠ 0. By [4, theorem 2.38, p.65], I = ann annI. Let J be a relative complement of I. So J is a minimal closed ideal of R, by Lemma 1.31. Hence J is a direct summand of R, since R is min-CS. Thus J = < e > for some idempotent e  R. But IJ  I  J = (0), so I  ann J = ann = <1 – e >. But <1 – e > is closed ideal of R, since it is a direct summand of R. It follows that I = <1 – e > because I is a maximal closed ideal of R. Thus I is a direct summand of R. Recall that, a ring R is called semihereditary if every finitely generating ideal of R is projective, [4, p.10]. 1.34 Corollary: Let R be a semiheriditary ring. Then R is a min-CS ring if and only if R is a max-CS ring. Proof: Since R is a semiheriditary ring. Then R is a nonsingular ring, [4, p.36]. So we get the result by Theorem 1.33. مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 Recall that a ring R is called regular (Von Neumann) if for every aR there is an x  R such that a x a = a, that is a = a2 x, [4, p.10]. 1.35 Corollary: If R is a regular ring. Then R is min-CS if and only if R is a max-CS. The following lemma is needed for the following corollary. 1.36 Lemma: [4, Exc.5, p.36] For a commutative ring, R/Z(R) is a nonsingular ring. Proof: Suppose there exists x = x + Z(R)  Z(R/Z(R)) and x  Z(R), so x ≠ 0 and ann(x) is not essential in R, also annR/Z(R)( x + Z(R))  e R/Z(R). Then for each I/Z(R) with I  Z(R), (I/Z(R)) annR/Z(R)( x+Z(R))≠OR/Z(R)=Z(R). So there exists a + Z(R)  ann R/Z(R)( x + Z(R)), a  I such that a  Z(R). Then a ≠ 0. Thus ax + Z(R) = Z(R), annR(a) is not essential in R and a  I. Hence ax  Z(R) and annR(a) is not essential in R. So that ann(ax)  e R and ann(a) is not essential in R. Now, since ann(ax)  e R, then for each nonzero ideal J of R, J  ann(ax) ≠ 0. Hence there exists y  ann(ax) and y ≠ 0. Thus yax = 0. (1) If ya = 0 then 0 ≠ y  ann(a) thus 0 ≠ y  Jann(a) so that ann(a)  e R which is a contradiction. (2) If ya ≠ 0, then 0≠yaann(x)J. Hence ann(x)  e R which is a contradiction. (3) If yx = 0 then 0 ≠ y  ann(x), so that 0 ≠ y  ann(x)  J. Thus ann(x)  e R which is a contradiction. (4) If yx ≠ 0, then 0≠ yxann(a)J, so that ann(a) e R which is a contradiction. (5) If ax = 0 then 0 ≠ a ann(x)  I, that is ann(x)  e R which is a contradiction. Thus our assumption is false and so Z(R/Z(R)) = Z(R) = OR/Z(R). 1.37 Corollary Let R be a ring. Then R/Z(R) is min-CS if and only if R/Z(R) is max-CS. Proof: By lemma 1.36 R/Z(R) is nonsingular, so the result follows by theorem 1.33. Before we give the following corollary, we need the following lemma. 1.38 Lemma: [4, Exc.13, p.37] If R is a nonsingular ring, then R[x1, x2, …, xn] is a nonsingular ring. Proof: First we shall prove that R[x] is a nonsingular. Let f(x)  Z(R[x]), then annR[x]f(x)  e R[x]. Assume f(x) = a0 + a1x + … + anx n, where ai  R, for each i = 0, 1, …, n. annR[x]f(x) = annR[x]a0  annR[x]a1x  …  annR[x]anx n  e R[x]. So, annR[x]a0 e R[x], annR[x]a1x  e R[x], …, annR[x]anx n  e R[x]. We claim that annR(ai)  e R, for all i = 0, 1, …, n. To prove this. Suppose there exists J  R and J ≠ 0 such that annRa0  J = 0. J ≠ 0 so J[x] is an ideal in R[x] and J[x] ≠ 0. Hence J[x]  annR[x]a0 ≠ 0. Let 0 ≠ g(x)  annR[x]a0 and g(x)  J[x], such that g(x) = b0 + b1x + … + bmxm, where bi  J, for all i = 0, 1, …, m and there exists k  {0,1,…,m} with bk ≠ 0. Since g(x)  annR[x](a), then (b0 + b1x + …+ bk x k +… + bmxm)a0 = 0. مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 Thus b0 a0 + b1 a0 x + …+ bk a0 x k + … + bm a0 x m = 0 which implies b0 a0 = 0, b1 a0 = 0, …, bk a0 = 0 , …, bm a0 = 0. But bk a0 = 0 and 0 ≠ bk  J implies that J  annRa0 ≠ 0, which is a contradiction with our assumption. Hence annRa0  eR, so that a0  Z(R). But R is nonsingular, hence a0 = 0. By the same way we can prove that each of a1,…, am belongs to Z(R)=(0). Thus f(x) = 0 and Z(R[x]) = (0); that R[x] is nonsingular. Then by induction R[x1, x2,…, xn] is a nonsingular ring. 1.39 Corollary: Let R be a nonsingular ring. Then R[x1, x2,…, xn] is min-CS if and only if R[x1, x2,…, xn] is a max-CS ring. Proof: By lemma 38 if R is a nonsingular ring, then R[x1, x2,…, xn] is a nonsingular ring. Hence the result follows by theorem 33. 1.40 Note: A direct sum of min (max)-CS modules will be discussed in another paper. References 1. Dung,N.V.; Huynh, D.V.; Smith, P.F. and Wisbauer, R., (1994), Extending Modules, John Wiley and Sons, Inc. New York. 2. Husain Suleman.S.Al-Hazmi, (2005), A Study of CS and -CS Rings and Modules, Ph.D.Thesis, College of Arts and Sciences of Ohio University . 3. Saad.H.M ohamed, Bruno J. Muller, (1990), Continuous and Discrete Modules, Combridge Univ. Press, New York, Port Chester Melbourne Sydney. 4. Goodearl, K.R. (1976), Ring Theory, Nonsingular Rings and Modules, Marcel Dekker, Inc. NewYork and Basel. 5. Mahmoud A. Kamal and Amany M. Menshawy, (2007), J. Egypt. Math.Soc., 15(2), . 157-168. 6. Kasch,F., (1982), M odules and Rings, Academic Press, Inc. London. 7. Z.A.EL-Bast and P.F.Smith, (1988), Multiplication Modules, Communication in Algebra, 10(4),.755-779. 8. C.P.Lu, (1984), Prime Submodules of Modules, Comment.M ath. Univ.St.Paul, 33, . 61-69. 9. Zeinab Talib Salman Al-Zubaidey, (2005), On Purely Extending Modules, M.Sc. Thesis, College of Science, University of Baghdad. 10. Ahmed, A.A. (1992), On Submodules of Multiplication Modules, M.Sc. Thesis, University of Baghdad. مقاسات التوسع) أعظم(أصغر مجلة إبن الھیثم للعلوم الصرفة و التطبیقیة 2012 السنة 25 المجلد 1 العدد Ibn Al-Haitham Journal for Pure and Applied Science No. 1 Vol. 25 Year 2012 ، رنا نوري مجید ھاديانعام محمد علي جامعة بغداد- ابن الھیثم -كلیة التربیة -قسم الریاضیات 2011 ایلول20:قبل البحث في ، 2011 أب 25:استلم البحث في الخالصة ن أصـغر مقـاس ) أعظـم( في ھذا البحث نعطي دراسة واسعة ألصغر ي المغلـق ـم اس الجزـئ ل المـق مقاسـات التوسـع مـث مقـاس توسـع ) أعظم(ومن بین النتائج االخرى نستعرض أنھ مركبة المجموع المباشر ألصغر . توسع ھو أصغر مقاس توسع ت الحلقـة ) أعظم(ھو أصغر ط إذا كانـت Rیـر مفـرده فـان الحلقـة غRمقاس توسـع وكـذلك اذا كاـن ـع إذا وفـق أعظـم حلقـة توس . أصغر حلقة توسع والتي ھي واحدة من المبرھنات المھمة في ھذا البحثRالحلقة . مقاس توسع، أصغر مقاس توسع، أعظم مقاس توسع، مقاس توسع منتظم:الكلمات المفتاحیة